AimQuantitative and transparent information on the ecological distribution niche of red panda (Ailurus fulgens) across Hindu-Kush Himalaya (HKH) region is still lacking. Our study presents the first quantitative large-scale prediction of the potential ecological niche of red panda for the entire HKH using publicly available ‘presence only’ data and best known scientific methods and techniques.
MethodsWe used 1120 red panda ‘presence only’ data collected from our own field works from 2007-2011, freely available bio-climatic variables (www.worldclim.org), and multiple machine learning algorithms such as Classification and Regression Trees, Random Forest, TreeNet, and Multivariate Adaptive Regression Splines to develop the models. We averaged all these models for the first produced Ensemble Model to assess the best area predictions of red panda for HKH.
ResultsOur predictive models have identified BIO4 (Temperature Seasonality) and BIO7 (Temperature Annual Range) are the major predictors for determining red panda ecological distribution niche in HKH. Based on Relative Index of Occurrence (ROI) with >0.3, our model estimates 56,000 km2 area as the global potential habitat of red panda in HKH and the majority of these highly suitable areas are still located outside the established protected areas of HKH region.
ConservationOur models can be used by wildlife managers to finding new red panda sub-populations, developing biological corridors between highly fragmented habitats to maintain gene flow and viable population in HKH region because the models are robust, transparent, publicly available, fit-for-use, and have a good accuracy, as judged by several metrics e.g. Receiver Operating Characteristics (ROC-AUC) curves, expert opinion and assessed by known absence locations.